Use your browser's BACK button to return to your page of origin.
A protocol for the analytical aspects of a systematic review of HIV prevention research.
Journal of Acquried Immune Deficiency Syndrome 2002;30(Suppl 1):S62-S72.
Johnson WD, Semaan S, Hedges LV, Ramirez G, Mullen PD, Sogolow E.
Abstract
Quantitative analysis can reveal the consistency of intervention effects across
studies, as well as the variation of effects according to study-level characteristics.
After consulting with project experts in methods and content, and reviewing
the literatures on research synthesis and on HIV prevention, we developed
a systematic protocol of analytical methods for synthesis of behavioral and
biologic outcome data from HIV intervention studies. This protocol included
procedures for identifying eligible studies; defining, characterizing, and
prioritizing outcomes; abstracting and calculating estimates of effect; adjusting
for baseline distributions and intraclass correlation; transforming estimates
to a common metric; summarizing effects; examining differences in effectiveness
among groups of studies; and translating these results into terms useful
to HIV prevention practitioners and researchers. We applied these procedures
to transform outcome data reported in many different statistical formats
into odds ratios that could be combined and compared across studies. We analyzed
data on behaviors related to sexual risk for HIV infection (unprotected sex,
condom use, and number of partners) as well as data on biologic outcomes
(incidence of HIV and other sexually transmitted infections). This framework
may be useful for meta-analyses of prevention research in other fields, particularly
when primary research features diverse outcome measures and methods of analysis.